PI 1.7.4 Problem during calibration

[tsaban]

Hi,

I appologize for taking your time. After installing 1.7.4 I have run into problems during calibration of my DSLR images. It seems when I add the bias calibration the calibrated image looks like being posterized (when auto stretched with STF). My bias, dark and flat images look quite as usual if I stretch them, also the subs look fine before calibrating them. I am attaching a screenshot that shows my Workspace.

I have converted my RAWs to FITS using 'Create RAW Bayer CFA...' producing mono images, but I have similar results when using 'Create RAW Bayer', the only difference is that they are colored  .

I am not sure if the problem happens because I did not back up my settings before installing 1.7.4 or if it is inherent to the new version. I currently have 1.7.4.0759 running on a Windows x64 machine. Any help will be much appreciated.

Thanks
Tahir Saban

[Nocturnal]

Hi,

it seems your frames don't match (different ISO settings perhaps) or perhaps your bias is getting subtracted twice.

[Cleon_Wells]

Question, did you recreate any of the Master Cal files with the new PI Ver?
What is your MasterFlat Scaling factor?
During light Cal is your Dark Scaling factor high, like 100.0 plus?
Try Calibrating your light with out the MasterDark.
Cleon

[tsaban]

Thanks Sander and Cleon,

it seems your frames don't match (different ISO settings perhaps) or perhaps your bias is getting subtracted twice.

@Sander: Yes, this is the case. The flat frames are ISO 800 while the rest is ISO 1600. However I did use different ISOs for flats before too without getting such an effect.

@Cleaon: I did create all the master calibration files with 1.7.4.

During light Cal is your Dark Scaling factor high, like 100.0 plus?
Try Calibrating your light with out the MasterDark.

Forgive me my ignorance, but I am not aware of the 'dark scaling factor'. I wonder how it can be checked?

When calibrating without the master dark the result seems to be ok (I mean without the posterization)

Regards
Tahir

[Nocturnal]

Hi,

if you do not have dark flats the your bias must match you flats:

(light - dark) / normalized(flat - dark flat)

[Cleon_Wells]

Click on the Process Console on the left side of the screen, you my think it's annoying/confusing at first glance but you can learn a lot from this information.
On the MasterDark, recreate it/integrate with uncalibrated Dark.fit sub files and check the Calibrate box in the MasterDark section. I have had issues with creating a MasterDark file from Calibrated Dark subs with my Canon T1i.
On the Math that Sander is talking about, I'll let him explain any Math question you have.
Cleon

[tsaban]

if you do not have dark flats the your bias must match you flats:

(light - dark) / normalized(flat - dark flat)

Aaarghhh, you got me!  Calibrating ISO800 flats with 1600 darks and biases can not be OK of course. There was also another problem. After the installation of 1.7.4 the output format in batch format conversion and other places was set to 'Same as input'. I have probably mixed it at other places withe the setting '32-bit float' since I used process icons saved with the old version. Anyway, after redoing the calibration (this time without calibrating the flats with the darks) the calibrated lights look good now. Thanks Sander and Cleon!

By the way, the reason for using a lower ISO for the flats was this: I make the flats in the field using an illuminated foil like this

http://www.teleskop-express.de/shop/product_info.php/language/en/info/p3866_Flatfield-Leuchtfolie-160-mm-im-Rahmen---Inverter---12V-Anschlus.html

If I use ISO 1600, the exposures are very short and I get a bad horizontal banding (probably caused by the 12V -> 220V inverter?). I am attaching a screenshot. Now I place a neutral grey foil between scope and illuminated foil and also reduce the ISO setting to lengthen the exposures. I have found that if the exposures are longer than 1/100s no banding is visible after the flats are stretched.

Did anyone of you expereince similar problems? If yes what do you do to solve it?

@Cleon: Thanks, I will check this.

[Nocturnal]

Cool!

Sorry, I don't use a DSLR.

[Cleon_Wells]

Tahir, I use an iso of 100 and a speed of 1/250 for my daylight/cloud sky Flat subs and the MasterFlat scaling is---, check the attached image of the Process Console window, Also note my strange Dark Scaling factors When I use calibrated Dark subs. I think the correct Dark Scaling factors are close to 1.0. Also note my MasterBias comes from 40 Bias subs   at iso100 and 1/4000 converted to fits with ether " Same as Target or 32 bit floating", after integration it's saved as 32bit Floating.  I can't see any difference between the two processes.
Cleon

[georg.viehoever]

...By the way, the reason for using a lower ISO for the flats was this: I make the flats in the field using an illuminated foil like this

http://www.teleskop-express.de/shop/product_info.php/language/en/info/p3866_Flatfield-Leuchtfolie-160-mm-im-Rahmen---Inverter---12V-Anschlus.html

If I use ISO 1600, the exposures are very short and I get a bad horizontal banding (probably caused by the 12V -> 220V inverter?). I am attaching a screenshot. Now I place a neutral grey foil between scope and illuminated foil and also reduce the ISO setting to lengthen the exposures. I have found that if the exposures are longer than 1/100s no banding is visible after the flats are stretched.
...

Yes, I see these as well. The solution is to take long exposure flats (>1 sec). Dim your foil by putting a couple of sheets of paper on top of it.
Georg

[tsaban]

Yes, I see these as well. The solution is to take long exposure flats (>1 sec). Dim your foil by putting a couple of sheets of paper on top of it.
Georg

I first tried paper, but when I heavily stretched the master flat, the non-uniform structure of paper became evident. Therefore I switched to a grey foil which is homogenuous. My understanding at the moment is that I have to apply more sheets of that foil to adequately dim the illumination.
Thanks
Tahir